A new realistic stochastic interpretation of Quantum Mechanics

[PeterDonis]

@Jacob Barandes welcome to Physics Forums!

 

[Greg Bernhardt]

I don't engage in technical discussions on internet forums.

Welcome to PF! This is a community you may want to reconsider. We have rather strict quality and civility guidelines.

 

[Jacob Barandes]

Oh, the civility and standards are great! That's clear from the thread! My firm rule about getting into technical discussions online is primarily a matter of time-management. I hope that's okay, and so sorry about this. –Jacob Barandes

P.S. And sorry in advance to those who email me for the slowness of my replies!

 

[Fra]

I finally skimmed through the video.

Reflections:

1) I think Aaronson's resistance is as exepcted for anyone juding an new paradigm from the perspective of another - as the way of reasoning is different. He wants to be convinced and see some explicit advantage, like what problems does the new paradigm solve. This is expected and fair enough.

Barandes is hinting at cosmoloigcal models and quantum gravity, and I personally buy that, and I presume that others that tuned into the idea, will understand it too, but I do not expect someone strongly rooted in a different paradigm of reasoning would buy into that without explicit results. Also fair enough.

2) But I think the heart of the matter, that is lifted but not resolved in the talk is:
the nature of law, nature of causality, and the role of observables, and the role of "non-observables".

These things are understood differently in different paradigms, which is I think at the root of the difficulty. After all, what is "causation" in the context of eternal timless laws and fixed state spaces? does the notion even make sense? because it's essentially timless?

A fine tuned hamiltonian is replaced by a fine tuned transition probability matrix. What differnce does it make one may ask?

Barandes seem to suggest that at some point where may be no "law", it's just stochastics? give the transition matrix. And that is then a key. I see that would allow a "stance" from where the usual fine tuning of dynamical law (via some extras not specified) can be emergent from stochastic processeses. So I personally think the potential promise I see is that this can help in unification and avoid fine tuning. This is why my key inquiry is, how does the transition matrix "evolve". This is then perhaps the closest thing to rules we get. It will not be a "dynamical law", but perhaps some "stochastic law" supplemented but evolutionary selection? This is they parts I am curious about.

3) I also like the discussion arond quantum computing, as it relates fo "computational effiency". If the "quantum representation" is more efficient, it might explain why non-commutative structiures are preferred by nature. Why does nature entertain "non-commutative" information? Is it simply more efficient from the perspective of encoding and computation? But that is not a "point" untile you actuall add some constraints of information processing capacity (beucause if you include limits of physicall processing, then time is lost) and the current QM paradigm cant construct these questions. And to connect to stochastics, an simple form of "computation" seems to be scrambling and black holes are sometimes thought of as the fastest scramblers in nature.

So there seems to be no lack of connections between stochastics, computing, and causality, and what is "observable". Even just thinking about this seems to require a new paradigm, as some likely key question can't even be constructed in the current paradigm and rejecting these questions may be a mistake.

/Fredrik

 

[pines-demon]

If anybody contacted Barandes (I haven't) and gets a response, please consider sharing his answer with us (ask him if you can anyway).

 

[martinbn]

Oh, the civility and standards are great! That's clear from the thread! My firm rule about getting into technical discussions online is primarily a matter of time-management. I hope that's okay, and so sorry about this. –Jacob Barandes

P.S. And sorry in advance to those who email me for the slowness of my replies!

So, you joined the forums to say that you are not going to participate!

 

[gentzen]

If anybody contacted Barandes (I haven't) and gets a response, please consider sharing his answer with us (ask him if you can anyway).

It is nice of you to be careful with Barandes's time. However, I guess you can still contact him later, if you wish.

So, you joined the forums to say that you are not going to participate!

Since Jacob joined the forums, he can share himself whatever parts of my email and his answers that he wish. Since he doesn't want to go into technical discussions online, I won't use such sharing as an excuse to start such a discussion. However, if the shared parts of my email should feel to selective to me, I reserve the right to share more of the surrounding context of the shared parts of my email.

 

[Fra]

So, you joined the forums to say that you are not going to participate!

It seems to me he rather took his time to join to tell us he is flattered and presumably appreciated the attention that his work is getting. I think that is nice and encouraging!

If he has no time to engage in some sometimes intense or lenghty discussions on here, that seems totally understandable. I presume he has many other channels to engage in as well, just as many others on here.

/Fredrik

 

[RedLocus]

I’ve been looking into Jacob Barandes’ stochastic-quantum correspondence, the idea of quantum systems as non-Markovian trajectories through a configuration space...well it's it’s compelling I must say, but something just doesn't sit well in my mind about quite a few points. One of which is how it fits in with more recent work on indefinite causal order, especially the process matrix framework developed by Brukner et al.

Barandes assumes a definite time parameter and fixed causal structure, while Brukner's work shows that causal order itself can be in superposition. Doesn’t this pose a major challenge for any theory grounded in stepwise, ordered trajectories? Is Barandes’ framework flexible enough to incorporate causal indefiniteness that seems to be being shown experimentaly? Or is this one of those conceptual threads where things might start to unravel?

 

[pines-demon]

I do not have enough time these days to watch 3h podcast videos, and I definitely was going to skip this new video by some unknown podcast featuring Barandes



However I could not wait to hear what he was going to say about Bell (interesting to hear, he has really matured on that, but I still think that I need to read the paper carefully). However I was not disappointed when I listened to that section because he mentioned us:

You know there's this discussion forum on Physics Forums that's been going on for for some time, and the people there are clearly very smart and have thought a lot about physics and quantum mechanics, and they've read these papers and they've seen some of these talks, and they're having discussions about things like "is this the right way to think about locality? or not the right way to think about locality?" And I I just love that they're doing this, because this is exactly the kind of work that we should all be doing. We should be thinking about how are we supposed to define things like locality and there are conversations about "is this picture compatible with remote entanglement swapping?" or different ways to think about about locality, and arguably it's compatible with all of these things.

You just have to carefully write it all out as a process, and treat everybody like atoms, and not have you know humans intervening, and you can't invoke measurements or collapse anymore because those aren't really happening, just play the whole thing out and just see it as it develops, and without being able to actually do projective measurements, there's not going to be the kinds of of strange non-locality that appears to be happening. But maybe I'm wrong and that's fine. But I think that the fact that people are actually discussing this is a good thing, and I hope people will do it more.

 

[physika]

causal indefiniteness that seems to be being shown <experimentaly>

Reference please

 

[javisot]

I understand that Barandes's claim that we should speak in a new language is because in this new stochastic and local paradigm entanglement and entanglement swapping do not exist, but it still explains all the experimental results equally. I would like to know the technical side of how this happens, how this new (local?) paradigm explains what in other contexts we call entanglement swapping (clearly non-local).

I also don't understand whether the proposal would correspond to standard QM, in the sense of offering the same but from a local perspective, or whether we go further (or don't reach that minimum).

 

[RedLocus]

Reference please

Indefinite Causal Order in a Quantum Switch

https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.121.090503

Experimental verification of an indefinite causal order

https://www.science.org/doi/10.1126/sciadv.1602589

Device-independent certification of indefinite causal order in the quantum switch

https://www.nature.com/articles/s41467-023-40162-8

 

[Fra]

Barandes assumes a definite time parameter and fixed causal structure

It's not how I would put his ideas?

As I understand Baranders the idea is rather that on the lowest level there is no strict causal laws, but simply guided stochastics. His new principle of causation is:

"A theory with microphysical directed conditional probabilities is causally local if any pair of localized systems Q and R that remain at spacelike separation for the duration of a given physical process do not exert causal influences on each other during that process, in the sense that the directed conditional probabilities for Q are independent of R, and vice versa."
-- https://arxiv.org/html/2402.16935v1#S5

Except for the subtle by critical detail where the notion of "spacelike" and division of the system into two Q and R; comes from in the reconstruction(*) this is IMO pretty much identical to how you would rationally think about causation also in terms of interacting randomly walking agents. Namely that the local map that guides the stochastic process, is independent of the maps of other agents. Loosely speaking I associate Q and R to inside agents. But "local" here I would in the general sense not refer to 4D spacetime metrics, but some abstract information space. Where local can simply mean, whatever one agent has at hand. "local information" would then be synonymous to "available information". And to construct spacelike relations between different kinds of information that are interacting, becomes a problem of the future. Indeed it is not possible to solve all at once. So when I understand it like this, I feel even more closer to his ideas that on first reading.

In his last youtube video it also becomes more to me that Barandes is not quite doing away with the observer - he is rather doing away with the EXTERNAL observers. And the only "observations" going inte those between parts of the system. That is a perspective that is exactly in line with how I think as well. That is another way to shifting towards the "inside observers" - and the point is that, that from this inside view, we have stochastic processes. The problem is not obsevers per see, but the fictional EXTERNAL observers. I have come to think that you can phrase things differently to make it look different, when it's not so different after all.

(*) Barandes also said in his last youtube clip in this thread that he has not yet and solution to quantum gravity. I presume that the subtle detail of putting in "spacelike" in that definition of causation, might need some subtle revision as well when we seek to incorporate gravity, and thus a evolving spacetime, so the notion of the distance metric might need generalization.

So to me, all this taken together suggests a picuture where causality is not fixed but emergent?

/Fredrik

 

[RedLocus]

It's not how I would put his ideas?

As I understand Baranders the idea is rather that on the lowest level there is no strict causal laws, but simply guided stochastics. His new principle of causation is:

"A theory with microphysical directed conditional probabilities is causally local if any pair of localized systems Q and R that remain at spacelike separation for the duration of a given physical process do not exert causal influences on each other during that process, in the sense that the directed conditional probabilities for Q are independent of R, and vice versa."
-- https://arxiv.org/html/2402.16935v1#S5

Except for the subtle by critical detail where the notion of "spacelike" and division of the system into two Q and R; comes from in the reconstruction(*) this is IMO pretty much identical to how you would rationally think about causation also in terms of interacting randomly walking agents. Namely that the local map that guides the stochastic process, is independent of the maps of other agents. Loosely speaking I associate Q and R to inside agents. But "local" here I would in the general sense not refer to 4D spacetime metrics, but some abstract information space. Where local can simply mean, whatever one agent has at hand. "local information" would then be synonymous to "available information". And to construct spacelike relations between different kinds of information that are interacting, becomes a problem of the future. Indeed it is not possible to solve all at once. So when I understand it like this, I feel even more closer to his ideas that on first reading.

In his last youtube video it also becomes more to me that Barandes is not quite doing away with the observer - he is rather doing away with the EXTERNAL observers. And the only "observations" going inte those between parts of the system. That is a perspective that is exactly in line with how I think as well. That is another way to shifting towards the "inside observers" - and the point is that, that from this inside view, we have stochastic processes. The problem is not obsevers per see, but the fictional EXTERNAL observers. I have come to think that you can phrase things differently to make it look different, when it's not so different after all.

(*) Barandes also said in his last youtube clip in this thread that he has not yet and solution to quantum gravity. I presume that the subtle detail of putting in "spacelike" in that definition of causation, might need some subtle revision as well when we seek to incorporate gravity, and thus a evolving spacetime, so the notion of the distance metric might need generalization.

So to me, all this taken together suggests a picuture where causality is not fixed but emergent?

/Fredrik

Your points about internal observers and emergent causation actually help quite a bit making some things here clearer. But still, the central issue remains the implicit assumption of a stable, fixed causal scaffolding in Barandes' framework. Even if causation emerges internally, his theory presumes some definite underlying linear causal structure. Experimentally, indefinite causal order is a thing, and explicitly challenges this assumption, showing scenarios where no single causal scaffolding can describe quantum processes. Bruckners process matrix is used to deal with these cases, but I fail to see how something like this can be used within Barandes' framework without reconceptualizations that would move it back to the wavefunction like territory. Anyway, I still see this as a fundamental tension that needs to addressed clearly.

 

[Fra]

central issue remains the implicit assumption of a stable, fixed causal scaffolding in Barandes' framework. Even if causation emerges internally, his theory presumes some definite underlying linear causal structure. Experimentally, indefinite causal order is a thing, and explicitly challenges this assumption, showing scenarios where no single causal scaffolding can describe quantum processes. Bruckners process matrix is used to deal with these cases, but I fail to see how something like this can be used within Barandes' framework without reconceptualizations that would move it back to the wavefunction like territory. Anyway, I still see this as a fundamental tension that needs to addressed clearly.

Ok I get your point and I agree it is a key issue.

My way of commenting this issue (from my stance) was in post 117:
"At the level of dynamics, the microphysical laws consist of conditional or transition probabilities of the form Γij(t) ≡ p(i, t|j, 0) [for i, j = 1, . . .N], (18) each of which supplies the probability for the system to be in its ith configuration at a continuously variable time t..."

Sounds reasonable and these obviously encode the corresponding hamiltonian details, but the question is, what is the process whereby these laws (transition probabilities) are inferred by a real observer. Without this, this seems to be out of taste for me. In principle I can imagine some elaborations where these transition amplities are constructed, but I see no traces of that in this thikning from skimming the papers. without this, this remains pursely descriptive, treating the "observer" as an implicit non-interacting context, just like most other interpretations."

I agree that just assuming a fixed transition matrix, as a starting point is no more satisfactorty than to assume a hilbert space and a hamilotonian. My main objection had to to with finetuning. IMO, I think this can be solved, but I agree it will take things to another level including reconceptualizations, that I think requires conconstructiing the "configuration space", which then includes spacetime. I am not sure if Barandes sees a "quicker way", but it's how I see if when judging his work from my perspective. So this will touch upon quantum gravity and I think also unification in general, which Baranders admits he has no solution yet.

In short, the way forward that seems natural to me, is to complement the stochastics in a fixed probability space, with a picture where the dimensionality and the probability space itself is evolving as well. I haven't seen Baranders mentioning anything like that, but if you relax everything at once, it becomes as huge mess. So assuming the fixes causal structure seems like a first step.

But this is where I think Barandes new view helps! Trying to imagine how probability spaces themselves evolve, is given all the learning models and AI that is popular now, is a lot easier, than to imagine a hilbert space evolving! Because we (or at least I don't) have a conceptual clear handle on hilbert space. In one of this talks he also highlighten that there are many things todo - and perhaps this new "picture" - will make some ideas more clear. I think the concept of evolutionary spaces here is an excellent example.

/Fredrik

 

[lodbrok]

Here is my current understanding of Barandes' work that I've not yet seen expressed anywhere: When we talk of a Stochastic process, we are not offering an ontology but rather an epistemology about how we model the system under consideration. Therefore, the mathematical tools we apply and the features of the mathematical model must be interpreted with great care to avoid misconceptions arising from ascribing ontology to mathematical concepts. For example, we can model Brownian motion as a stochastic process. However, we could also use a different mathematical model with trajectories and interactions for individual particles. It won't be very easy, but it is possible in principle. Therefore, we have to be careful not to think that the model's features are necessarily features of the underlying system.

Looking a bit more broadly, we can model a classical system using Newtonian mechanics, a Lagrangian, or a Hamiltonian. Similarly, we have to be careful when we interpret the features of the mathematical tool as features of the underlying system. This is one of the reasons Barandes complains about (and I agree) approaches that ascribe ontology to the wavefunction or conclude from the path-integral approach that particles are taking all possible paths, etc.

The advantage of Barandes' approach is that he starts with an assumption that an underlying system exists; he makes no claims about what the system is actually doing. But then he proceeded to show that if he models the system as an "indivisible stochastic process", he can recover all the equations of quantum mechanics. The stochasticity and the indivisibility are features of the model, not features of the underlying system. This gives a conceptually clear distinction between the map and the territory, allowing us to ask questions about the underlying system in different ways. Some questions would involve extending the model in "model-space" to see what more it can do. Still, many other questions would involve looking more closely at the underlying system to see what additional models could be offered to extend our understanding. Unification attempts with general relativity could then be approached both ways.

This is why when Scott Aaronson asks: "Where are the trajectories?", it appears to me he probably hasn't understood the model. The trajectories belong to the underlying system not the current model. I see a lot of discussion in which the distinction between the model and the system is not made clear enough. Barandes himself could make this point more clear. For example, I don't believe he has explicitly stated that it is not the underlying system that is "indivisible" but rather the process used to model it.

 

[PeterDonis]

The trajectories belong to the underlying system not the current model.

We don't know that trajectories belong to the underlying system. Trajectories themselves are a model--a model carried over from classical mechanics.

 

[Fra]

When we talk of a Stochastic process, we are not offering an ontology but rather an epistemology about how we model the system under consideration. Therefore, the mathematical tools we apply and the features of the mathematical model must be interpreted with great care to avoid misconceptions arising from ascribing ontology to mathematical concepts.
...
Looking a bit more broadly, we can model a classical system using Newtonian mechanics, a Lagrangian, or a Hamiltonian. Similarly, we have to be careful when we interpret the features of the mathematical tool as features of the underlying system. This is one of the reasons Barandes complains about (and I agree) approaches that ascribe ontology to the wavefunction or conclude from the path-integral approach that particles are taking all possible paths, etc.

The advantage of Barandes' approach is that he starts with an assumption that an underlying system exists; he makes no claims about what the system is actually doing. But then he proceeded to show that if he models the system as an "indivisible stochastic process", he can recover all the equations of quantum mechanics. The stochasticity and the indivisibility are features of the model, not features of the underlying system. This gives a conceptually clear distinction between the map and the territory, allowing us to ask questions about the underlying system in different ways. Some questions would involve extending the model in "model-space" to see what more it can do. Still, many other questions would involve looking more closely at the underlying system to see what additional models could be offered to extend our understanding. Unification attempts with general relativity could then be approached both ways.

If you mean what I think you mean(?), then I agree. Here is how i view what you say, let me know if i misinterpreted something.

But at the same time as we emhpasise the difference between ontology and epistemology, the whole approach deeply suggests that they are connected or related.

They key is to distinguish between the conventional epistemology from the perspective of the external observer, including the human science context; and the epistemological perspective of the "internal observers" as in small parts of the system interacting with other parts.

The former perspective can easily be dismissed as not beeing relevant to nature, but the latter perspective (if we can make sense of it) really mixes "epistemology" and "measurement" or physical inferences, as in how parts relate or are aware of other parts in it's own physical environment.

So, even if we don't know - it is suggestive that if the parts behaves "as if" then are stochastically evolving as per the Baranders causally local directed probabilities, that may tell us something about the ontology. It is how I see it, and it is one of the key insights.

And this can be imagine to be expanded into "model-space" in future work, if that's what you mean we are in agreement.

I think the reason why we confuse the models suggestion of taking all paths at once, is because that is what it looks like from the perspective of the external observer.

From the inside however, it does not talk all paths, it takes one, but stochastically so. But if the parts are all doing this in a locally causal way, it means that the premises of bells theorem are not valid. The indivisibility and causal locally directed probabilities simply can not be causally understood from the perspective of the external observer; it only seems "quantum weirdness". But if we consider the inside perspective, this weirdness seems to be gone. The weirdness is an artefact from trying to infer causal patterns from correlations, as that as all we have from the external view.

This is why when Scott Aaronson asks: "Where are the trajectories?", it appears to me he probably hasn't understood the model. The trajectories belong to the underlying system not the current model.

For sure the trajectories are not "observables" in the traditional sense, nor can they be inferred by an arbitrary observers. Thats how I think they are "real" as in actually there - but hidden to the extent that it even escapes causally influencing other parts of the system!

/Fredrik

 

[lodbrok]

They key is to distinguish between the conventional epistemology from the perspective of the external observer, including the human science context; and the epistemological perspective of the "internal observers" as in small parts of the system interacting with other parts.

The former perspective can easily be dismissed as not beeing relevant to nature, but the latter perspective (if we can make sense of it) really mixes "epistemology" and "measurement" or physical inferences, as in how parts relate or are aware of other parts in it's own physical environment.

I agree with this. To me the latter perspective would be closer to ontology. Here is what Barandes says in his paper on Causal locality. What he calls nomology, I would have considered as a form of epistemology.
https://arxiv.org/html/2402.16935v1

In the context of this overall stochastic picture, the wave function is not a piece of ontological furniture, but instead encodes epistemic information—the system’s probabilities—as well as nomological information—the system’s unistochastic microphysical dynamics.

So, even if we don't know - it is suggestive that if the parts behaves "as if" then are stochastically evolving as per the Baranders causally local directed probabilities, that may tell us something about the ontology. It is how I see it, and it is one of the key insights.

Agree 100%. This to me is the key also.

 

[iste]

Indefinite Causal Order in a Quantum Switch

https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.121.090503

Experimental verification of an indefinite causal order

https://www.science.org/doi/10.1126/sciadv.1602589

Device-independent certification of indefinite causal order in the quantum switch

https://www.nature.com/articles/s41467-023-40162-8

I’ve been looking into Jacob Barandes’ stochastic-quantum correspondence, the idea of quantum systems as non-Markovian trajectories through a configuration space...well it's it’s compelling I must say, but something just doesn't sit well in my mind about quite a few points. One of which is how it fits in with more recent work on indefinite causal order, especially the process matrix framework developed by Brukner et al.

Barandes assumes a definite time parameter and fixed causal structure, while Brukner's work shows that causal order itself can be in superposition. Doesn’t this pose a major challenge for any theory grounded in stepwise, ordered trajectories? Is Barandes’ framework flexible enough to incorporate causal indefiniteness that seems to be being shown experimentaly? Or is this one of those conceptual threads where things might start to unravel?

Looking at some of the papers on this topic by Brukner et al., seems to me this stuff can be seen as another form of contextuality that I don't think would be problematic since contextuality is more or less the essence of quantum mechanics which any formulation or interpretation will need to cover.

There is a long list of things that have indefiniteness in quantum mechanics, which all seem describable in the same way related to what is basically contextuality. This is just one example, Wigner's friend is another - and I believe there are parallels between this work and Brukner's work on Wigner's friend invoking joint probability distributions. Barandes' approach emphasizes yet another example of indefiniteness in terms of the indefiniteness of trajectories - not all of the joint probability distributions regarding trajectories can be constructed in the theory (which is just a characteristic of quantum theory) - while you can also see explicitly in his papers interference, non-commutativity, the uncertainty principle which are all yet again examples of this same kind of contextual behavior.

Whenever these claims of indefiniteness come about, they all seem characterizable in basically the same kind of way in terms of statistical descriptions so in principle this causal indefiniteness is not problematic for Barandes' kind of formulation and its not fundamentally different to the other kinds of indefiniteness you already see in his formulation and quantum theory generally. I would wager that there is any inherent need to alter Barandes' model to accommodate this stuff and it is not any more remarkable than things like Bell violations and other forms of contextuality which I would not expect a problem with.

I understand that Barandes's claim that we should speak in a new language is because in this new stochastic and local paradigm entanglement and entanglement swapping do not exist, but it still explains all the experimental results equally. I would like to know the technical side of how this happens, how this new (local?) paradigm explains what in other contexts we call entanglement swapping (clearly non-local).

Entanglement does exist in the Barandes formulation and he describes it in terms of indivisible stochastic systems.

If two systems interact (locally), this can induce correlations between the systems rendering them non-separable / non-factorizable. If the systems are Markovian, then once they stop interacting the correlation disappears because the system behavior only depends on the current state: no interaction, no correlation. For indivisible systems, any statistical information all the way down to the initial time will continue to be stored cumulatively in the system's transition matrices so that even when the two systems are not interacting, the correlation induced by the initial local interaction is still preserved in the behavior of the stochastic system - it does not only depend on the current non-interacting state.

All the elements required for entanglement swapping in terms of non-separability and division events exist in the Barandes formulation so it can be kind of envisioned how entanglement swapping could be expressed in indivisible systems. Local interactions between two separate systems create a non-separable composite system. If a single sub-system (e.g. a1) of one non-separable composite system A interacts with a single sub-system (e.g. b1) of another non-separable composite system B, the two systems A and B will no longer be non-separable but the interacting sub-components (i.e. a1 and b1) will now be non-separable. That is in a very basic way describing the swap event in entanglement swapping.

I also don't understand whether the proposal would correspond to standard QM, in the sense of offering the same but from a local perspective, or whether we go further (or don't reach that minimum).

At the center of his formulation is a theorem demonstrating that any unitary quantum system can be expressed as an indivisible one and vice versa so its showing that in principle for every kind of standard QM scenario you can talk about, there is an equivalent description in terms of an indivisible stochastic system.

I think maybe one way of looking at the locality issue - but only as a stop-gap - is that rather than explicitly saying it is local (which is a very ambiguous word and can mean different things that may require more assumptions than Barandes can offer), we can almost definitively say that there are no explicit non-local influences in the mathematics of the model, particularly of the kind that would require the kind of foliations you would eventually need in Bohmian mechanics due to quantum potential. Systems don't spontaneously affect each other across distances beyond the preservation of an initial locally-induced correlation, measurements of which do not result in causal interactions in a similar sense to non-signalling. This is also the case in standard QM as well as long as there is no physical collapse.

We don't know that trajectories belong to the underlying system. Trajectories themselves are a model--a model carried over from classical mechanics.

But they do in Barandes' formulation and, as lodbrok says, indivisibility seems to be sufficient for a system with trajectories to produce quantum behavior, continuous trajectories or otherwise (presumably)

 

[javisot]

Looking at some of the papers on this topic by Brukner et al., seems to me this stuff can be seen as another form of contextuality that I don't think would be problematic since contextuality is more or less the essence of quantum mechanics which any formulation or interpretation will need to cover.

There is a long list of things that are indefiniteness quantum mechanics which all seem describable in the same way related to what is basically contextuality. This is just one example, Wigner's friend is another - and I believe there are parallels between this work and Brukner's work on Wigner's friend invoking joint probability distributions. Barandes' approach emphasizes yet another example of indefiniteness in terms of the indefiniteness of trajectories - not all of the joint probability distributions regarding trajectories can be constructed in the theory (which is just a characteristic of quantum theory) - while you can also see explicitly in his papers interference, non-commutativity, the uncertainty principle which are all yet again examples of this same kind of contextual behavior.

Whenever these claims of indefiniteness come about, they all seem characterizable in basically the same kind of way in terms of statistical descriptions so in principle this causal indefiniteness is not problematic for Barandes' kind of formulation and its not fundamentally different to the other kinds of indefiniteness you already see in his formulation and quantum theory generally. I would wager that there is any inherent need to alter Barandes' model to accommodate this stuff and it is not any more remarkable than things like Bell violations and other forms of contextuality which I would not expect a problem with.



Entanglement does exist in the Barandes formulation and he describes it in terms of indivisible stochastic systems.

If two systems interact (locally), this can induce correlations between the systems rendering them non-separable / non-factorizable. If the systems are Markovian, then once they stop interacting the correlation disappears because the system behavior only depends on the current state: no interaction, no correlation. For indivisible systems, any statistical information all the way down to the initial time will continue to be stored cumulatively in the system's transition matrices so that even when the two systems are not interacting, the correlation induced by the initial local interaction is still preserved in the behavior of the stochastic system - it does not only depend on the current non-interacting state.

All the elements required for entanglement swapping in terms of non-separability and division events exist in the Barandes formulation so it can be kind of envisioned how entanglement swapping could be expressed in indivisible systems. Local interactions between two separate systems create a non-separable composite system. If a single sub-system (e.g. a1) of one non-separable composite system A interacts with a single sub-system (e.g. b1) of another non-separable composite system B, the two systems A and B will no longer be non-separable but the interacting sub-components (i.e. a1 and b1) will now be non-separable. That is in a very basic way describing the swap event in entanglement swapping.


At the center of his formulation is a theorem demonstrating that any unitary quantum system can be expressed as an indivisible one and vice versa so its showing that in principle for every kind of standard QM scenario you can talk about, there is an equivalent description in terms of an indivisible stochastic system.

I think maybe one way of looking at the locality issue - but only as a stop-gap - is that rather than explicitly saying it is local (which is a very ambiguous word and can mean different things that may require more assumptions than Barandes can offer), we can almost definitively say that there are no explicit non-local influences in the mathematics of the model, particularly of the kind that would require the kind of foliations you would eventually need in Bohmian mechanics due to quantum potential. Systems don't spontaneously affect each other across distances beyond the preservation of an initial locally-induced correlation, measurements of which do not result in causal interactions in a similar sense to non-signalling. This is also the case in standard QM as well as long as there is no physical collapse.



But they do in Barandes' formulation and, as lodbrok says, indivisibility seems to be sufficient for a system with trajectories to produce quantum behavior, continuous trajectories or otherwise (presumably)

Great answer, I see the following paragraphs (1,2) contradictory respect to the last paragraph (3)

(1)-"Entanglement does exist in the Barandes formulation and he describes it in terms of indivisible stochastic systems."

(2)-"All the elements required for entanglement swapping in terms of non-separability and division events exist in the Barandes formulation so it can be kind of envisioned how entanglement swapping could be expressed in indivisible systems. "


(3)-"we can almost definitively say that there are no explicit non-local influences in the mathematics of the model, particularly of the kind that would require the kind of foliations you would eventually need in Bohmian mechanics due to quantum potential. Systems don't spontaneously affect each other across distances beyond the preservation of an initial locally-induced correlation, measurements of which do not result in causal interactions in a similar sense to non-signalling."


What we understand by entanglement and entanglement swapping is typically explained using the existence of non-local influence. A local theory is a theory that does not contain non-local influences, obviously. By this I mean that, in Barandes's formulation, these phenomena (entanglement and entanglement swapping) are as local and common as any other phenomenon in that paradigm.

I have one question left:
Is there any calculation that, using this new paradigm, is easier (or harder) than if I perform the same calculation using a different standard paradigm? Does this allow us to get anywhere more efficiently?

 

[DrChinese]

All the elements required for entanglement swapping in terms of non-separability and division events exist in the Barandes formulation so it can be kind of envisioned how entanglement swapping could be expressed in indivisible systems. Local interactions between two separate systems create a non-separable composite system. If a single sub-system (e.g. a1) of one non-separable composite system A interacts with a single sub-system (e.g. b1) of another non-separable composite system B, the two systems A and B will no longer be non-separable but the interacting sub-components (i.e. a1 and b1) will now be non-separable. That is in a very basic way describing the swap event in entanglement swapping.

Thanks for at least mentioning Entanglement Swapping.

I guess you would then say that if there were a third, fourth, etc composite systems C/D/etc between systems A and B: then the same effect could cascade through such intermediate systems to leave a1 and b1 to have those “spurious” perfect correlations. (Since entanglement isn’t real in this view).

But now you have these nasty issues with causality: because some subsystems can have swaps both before and/or after other subsystems cease to exist. Systems A and B can completely cease to exist before the connecting systems are even created. This is in fact the basis for quantum repeaters.

 

[iste]

What we understand by entanglement and entanglement swapping is typically explained using the existence of non-local influence. A local theory is a theory that does not contain non-local influences, obviously. By this I mean that, in Barandes's formulation, these phenomena (entanglement and entanglement swapping) are as local and common as any other phenomenon in that paradigm.

Yes but this is the ambiguous aspect that I was trying to avoid. There will are non-local correlations in Barandes' theory just like quantum theory, but it is very difficult to say what these correlations mean or how they come about. Certainly, I don't think you can do that without going beyond what Barandes' formalism and into further interpretation.

But since we cannot resolve that right now, we can say at the very least that there is nothing explicit in the math where you can do something to a system, or something happens to it, which has an instantaneous effect somewhere else in the theory - at least not in what Barandes has showed so far...

I guess if there were such explicit instantaneous influences then someone could still argue something like "its only epistemic" but then you would need a good way to explain it away. So at least lack of explicit instantaneous influence means one less argument you would need, and you could say the theory is no more non-local than orthodox quantum theory. In contrast, things like Bohmian mechanics or the Markovian Nelsonian version of stochastic mechanics (but not the non-Markovian Levy-Krener version) do have this extra kind of feature which makes it very explicit-in-your-face-directly in the mathematics of the theory that doing something to some particle can have an instantaneous effect on another, because its basically built into things like the velocities and Bohmian potential. That kind of thing doesn't happen in the Barandes picture. Its then not clear that you can look at the Barandes theory less favorably than normal quantum theory in terms of non-locality.

Yeah, you can argue about deeper interpetations of entanglement in the theory and whether it truly does or does not represent some local or non-local force. But that requires further speculation. At the very least, if you want some position to stand on now, one can say that explicit, instantaneous non-local influences don't appear to exist like they do for Bohm or Nelson. I would then say a position is on a bit more of stronger standing when accusations of non-local causation come from the personal incredulity of the accuser as opposed to it being explicily written in the theory like with Bohmian or Nelsonian mechanics. At least then the position is more or less at the same place as the orthodox quantum theory without needing an additional thing to explain away (and arguably people in orthodox theory still need to explain how collapse doesn't lead to instantaneous influence whereas in Barandes' theory it doesn't exist beyond Bayesian statistical conditioning which is explicitly not physical).

So my statements about non-locality in what you replied to were supposed to be a kind of compromise for the fact that there - at least in my opinion - there is no smoking gun here where you can unequivocally say local or non-local, metaphysically speaking. I think that would need deeper elaboration about what the theory says is happening and why exactly that happens, especially in certain scenarios like spin correlations. I don't think orthodox quantum theory or other interpretations have a smoking gun either though. You say we explain entanglement through non-local influences; but afaik, this consensus comes from Bell, and I don't think Bell violations explicitly prove non-local influences. What they prove is that a noncontextual model cannot explain the data; nonlocal influences is an intuitive way of filling the vacuum left discrediting noncontextual models, but I don't think Bell violations explicitly prove that nonlocal influences must be what fills the vacuum.

It would be also interesting though to see what would happen if you were to model the specific bell-state entanglement swapping scenarios in the Barandes theory, especially with regard to the Mjelva discussion in another recent thread - since entanglement swapping does prima-facie seem more radical. But I guess, that model would require more elaborations in the Barandes theory to be made beforehand.

Is there any calculation that, using this new paradigm, is easier (or harder) than if I perform the same calculation using a different standard paradigm? Does this allow us to get anywhere more efficiently?

I don't know. Barandes talks about how the fact that the unitary representation of a system being divisible makes it more efficient compared to looking through the stochastic representation of the same system which is indivisible. Presumably thats how any advantages in quantum computing would come about according to Barandes formulation.

 

[iste]

Thanks for at least mentioning Entanglement Swapping.

I guess you would then say that if there were a third, fourth, etc composite systems C/D/etc between systems A and B: then the same effect could cascade through such intermediate systems to leave a1 and b1 to have those “spurious” perfect correlations. (Since entanglement isn’t real in this view).

But now you have these nasty issues with causality: because some subsystems can have swaps both before and/or after other subsystems cease to exist. Systems A and B can completely cease to exist before the connecting systems are even created. This is in fact the basis for quantum repeaters.

Well entanglement is real in the Barandes view because its just saying a quantum system provably is an indivisible stochastic one which carries all the same properties including the ones that belong to entanglement. If the quantum system doesn't have a problem with measurement and swap orders then I should think that the Barandes one certainly wouldn't given that in principle the quantum system could be translated into the Barandesian one. Neither is it really clear to me why measurement orders should matter or have consequences for causality for entanglement generally in this formulation when Barandes' explanation of entanglement seems to be on the level of correlations preserved over time without measurements explicitly having additional disturbing causal effects on spatially distant systems. In order to explain why the swap order doesn't matter for Barandes' formulation, I imagine something along the lines of the Mjelva post-selection would have to be used.

[Edited for clarity]

 

[physika]

it is not the underlying system that is "indivisible" but rather the process

Interesting, the fundamental is not the "system" or the "event", PROCESS is the fundamental.

Something like;

"We thus choose to determine necessary conditions that are operationally useful in identifying or falsifying causality claims. Our proposed approach is based on stochastics, in which events are replaced by PROCESSES"

Revisiting causality using stochastics.
https://royalsocietypublishing.org/doi/10.1098/rspa.2021.0835

​

 

[lodbrok]

Interesting, the fundamental is not the "system" or the "event", PROCESS is the fundamental.

Something like;

"We thus choose to determine necessary conditions that are operationally useful in identifying or falsifying causality claims. Our proposed approach is based on stochastics, in which events are replaced by PROCESSES"

Revisiting causality using stochastics.
https://royalsocietypublishing.org/doi/10.1098/rspa.2021.0835

​

These authors may have a different view to Barandes as to the meaning of fundamental. He addresses this in the first 30 minutes of the most recent video posted above. In any case, when developing a theory, you choose what is fundamental "according to the theory".

I wonder, how would you define "PROCESS" without (explicitly or implicitly) making use of the concepts of a "system" or "event"?

 

[physika]

These authors may have a different view

Plausible.

 

[Fra]

But now you have these nasty issues with causality: because some subsystems can have swaps both before and/or after other subsystems cease to exist. Systems A and B can completely cease to exist before the connecting systems are even created. This is in fact the basis for quantum repeaters.

I'll make an attempt to paint a picture on how I would conceptually describe the bell entanglement experiment using how I understand Barandes perspective:

In Barandes picture I would say that entanglement between two quantum systems can be understood as a pre-correlation between two independent separated stochastic processes. But this correlation is valid only as long as they two matrices aren't updated. (This is attained by the isolation that also in the standard picture prevents decoherence.)

But there is NO causal relation between any future evolution of these entangled system. The causal connection is only in the pre-tuning of their "predictive encoding".

But the pre-correlated transition matrices, does not mean that any actual outcomes of measurements are pre-determined, because these are really an interaction between the stochastic system; and the environment - the "apparatous settings" (to avoid unnecessary details) and the apparatous can not be conditioned on anything but the available information - which is not the transition matrices, but what is at hand in the macroscopic enviromnent, which is the "preparation procedure".

So the "hidden variable" that pre-correlates the two systems guided stochastics does not determined the outcomes; it only introduces a bias in the outcomes. The acutal outcome also depends on the detector and it's settings, and these are not part of the original stochastic matrix.

So conceptualy we as Baranders also suggests, a kind of "hidden variable explanation", but not a the traditional one that bell rules out. But one where it's the probability matrices of that are "predetermined" - not the full outcomes when interacting with the measurement device.

So we need no action at distance. It all makes perfect conceptual sense.

What is taking the huge blow here is IMO our previous understanding of "causality". I'd put it like this: An external observer can NEVER "observe" causal pattern between parts. It only observes correlations. The only "causal" notion, is the "guide" to the stochastics. We never ge closer to the facts. But as this population of stochastic process evolve, effective laws and effective causal laws may be emergent. This is how I also interpret some of Baranders thinking after listening to his clips. And I am all onboard with this.

You can consider this all handwaving, and indeed what is admittedly MISSING still, is the full updated theory that explicitly models how these "matrices" evolve, and how a network of them is consistent with or even allows for an emergent 4D spacetime. Right now the way they evolve are not explain a priori by a new theory, they would have to be explained via regular hilbert picture; and then via transformations; as all we have atm is the correspondence; we are still awaiting the full liberation, and i think that will require quite a reconstruction that likely includes quantum gravity and other forces.

I think this also relates to the issueo of RedLocus in post 585.

But I think the perspective suggested by the correspondence, is a more intuitive platform, than the alternavite as it offers a better handle on the inside view, required to get rid of the external fictional observers.

/Fredrik

 

[DrChinese]

I'll make an attempt to paint a picture on how I would conceptually describe the bell entanglement experiment using how I understand Barandes perspective:

...

1. But there is NO causal relation between any future evolution of these entangled system.
2. The causal connection is only in the pre-tuning of their "predictive encoding".

Assuming you were tying your comments to my comment about quantum repeaters (that utilize entanglement swapping):

1. This is contradicted by experiment. An experimenter's future free choice can create a swap from previously independent systems.

2. This is his hypothesis, I guess. Independently created entangled pairs (indivisible systems) cannot, according to quantum theory, have any "pre-tuning". There is no such thing. All such pairs are in the same superposition (Bell State). So if the hypothesis were true, that should show up in experiment - and it would violate Monogamy relations (i.e. theory). So this hypothesis is a huge leap.